U.S. patent application number 17/525461 was filed with the patent office on 2022-05-12 for respiratory compensated robot for liver cancer treatment.
This patent application is currently assigned to BOARD OF TRUSTEES OF THE UNIVERSITY OF ARKANSAS. The applicant listed for this patent is BOARD OF TRUSTEES OF THE UNIVERSITY OF ARKANSAS. Invention is credited to Yue Chen, Mishek Musa.
Application Number | 20220142702 17/525461 |
Document ID | / |
Family ID | 1000005996516 |
Filed Date | 2022-05-12 |
United States Patent
Application |
20220142702 |
Kind Code |
A1 |
Chen; Yue ; et al. |
May 12, 2022 |
Respiratory Compensated Robot for Liver Cancer Treatment
Abstract
A robotic platform system having a lower stage with a motorized
cartesian carriage, an upper stage, and a needle insertion module
that connects both stages together.
Inventors: |
Chen; Yue; (Fayetteville,
AR) ; Musa; Mishek; (Fayetteville, AR) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
BOARD OF TRUSTEES OF THE UNIVERSITY OF ARKANSAS |
Fayetteville |
AR |
US |
|
|
Assignee: |
BOARD OF TRUSTEES OF THE UNIVERSITY
OF ARKANSAS
Fayetteville
AR
|
Family ID: |
1000005996516 |
Appl. No.: |
17/525461 |
Filed: |
November 12, 2021 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
63113050 |
Nov 12, 2020 |
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
A61B 2017/00557
20130101; A61B 2018/00577 20130101; A61B 18/1477 20130101; A61B
2018/00184 20130101; A61B 34/30 20160201; A61B 2017/00398 20130101;
A61B 2018/00529 20130101 |
International
Class: |
A61B 18/14 20060101
A61B018/14 |
Claims
1. A robotic platform system that enables accurate needle
deployment in the dynamic environment of the liver comprising: (1)
a lower stage with a motorized cartesian carriage, (2) an upper
stage with a motorized cartesian carriage, and (3) a needle
insertion module that connects both stages together.
2. The system of claim 1 wherein the upper and lower stages have
carriages that can move in both the x and y directions.
3. The system of claim 2 wherein the carriages both have spherical
bearings set into them that support said needle insertion module
and by changing the relative location between the bearings, the
orientation of the needle insertion module can be controlled about
the x and y axes.
4. The system of claim 3 wherein said needle insertion module
includes a flexible fluidic actuator adapted to insert and retract
a needle.
5. The system of claim 4 wherein said flexible fluidic actuator
includes one or more inflatable bellows adapted to move a needle
linearly when said bellows are inflated.
6. The system of claim 5 wherein said flexible fluidic actuator
includes one or more diaphragms that hold a needle stationary
within said robot when inflated.
7. The system of claim 6 wherein, when a needle is to be inserted
further into a patient, said one or more diaphragms are inflated
prior to said inflation of said bellows.
8. The system of claim 7 wherein said upper and lower stages are
identical.
9. The system of claim 8 further including a linear optical encoder
and a linear transmissive strip, said optical encoder and linear
transmissive strip adapted to provide the relative displacement od
said FFA.
10. The system of claim 7 further including a gap between said
needle and said diaphragm and bellows when said diaphragms and
bellows are deflated.
11. The system of claim 10 wherein said gap eliminates contact
between said needle and said bellows and said diaphragm when said
bellows and said diaphragm are deflated.
12. The system of claim 10 wherein said needle does not contact
said bellows and said diaphragm when said bellows and said
diaphragm are deflated.
13. A method for accurate needle deployment in the dynamic
environment of the liver comprising the steps of: providing a
robotic platform having (1) a lower stage with a motorized
cartesian carriage, (2) an upper stage, and (3) a needle insertion
module that connects both stages together; said upper and lower
stages have carriages that can move in both the x and y directions;
said insertion module includes of a fluidic actuator adapted to
insert and retract a needle; said flexible fluidic actuator
includes one or more inflatable bellows adapted to move a needle
linearly when said bellows are inflated and one or more diaphragms
that hold a needle stationary within said robot when inflated; when
a needle is to be inserted during a static phase of a liver, said
one or more diaphragms are inflated to grip the needle, then said
one or more bellows are inflated to insert the needle; and prior to
a liver transitioning out of a static phase, said one or more
inflatable diaphragms are deflated and then said one or more
bellows are deflated.
14. The method of claim 13 wherein said needle is allowed to move
freely without interference within said fluid actuator when said
needle is not being inserted.
15. The method of claim 13 further including a linear optical
encoder and a linear transmissive strip, said optical encoder and
linear transmissive strip adapted to provide the relative
displacement of said FFA.
16. The method of claim 13 further including a gap between said
needle and said diaphragm and bellows when said diaphragms and
bellows are deflated.
17. The method of claim 16 wherein said gap eliminates contact
between said needle and said bellows and said diaphragm when said
bellows and said diaphragm are deflated.
18. The method of claim 13 wherein said needle does not contact
said bellows and said diaphragm when said bellows and said
diaphragm are deflated.
19. The method of claim 13 wherein said needle is inserted for
around 2 seconds.
20. The method of claim 13 wherein said upper and lower stages are
identical.
Description
RELATED APPLICATIONS
[0001] This application claims priority to U.S. Provisional
Application No. 63/113,050, filed on Nov. 12, 2020, which is
incorporated herein in its entirety.
STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH &
DEVELOPMENT
[0002] Not applicable.
INCORPORATION BY REFERENCE OF MATERIAL SUBMITTED ON A COMPACT
DISC
[0003] Not applicable.
BACKGROUND OF THE INVENTION
[0004] Primary liver cancer, also known as hepatocellular carcinoma
(HCC), is the third most common cause of cancer-related death in
the world with over 700,000 deaths reported annually. In the United
States, an estimated 42,000 adults are diagnosed with liver cancer
each year, and the number of cases is expected to continuously rise
due to the increasing number of chronic liver diseases caused by
alcohol, nonalcoholic fatty liver disease, hepatitis B, and
hepatitis C infection. The five-year survival rate for patients is
as low as 4%. As a result, there is a considerable economic loss of
upwards of $1 billion per year in the United States. HCC can be
treated with a variety of methods. Medical therapies, such as
sorafenib, provides no reduction to the mortality rate as it only
prolongs survival for a few months. Chemotherapy fails to provide
effective treatment to control tumor growth, primarily due to HCC's
resistance to radiation. Liver transplantation and partial surgical
resection have both been shown to be effective methods of
treatment, however, they both require strict criteria for candidate
selection, and this prevents the majority of diagnosed patients
from receiving treatment. Stereotactic radiosurgery (SRS) is a
feasible approach for patients who are not eligible for liver
transplantation or surgical resection; however, randomized clinical
trials are still needed to justify its effectiveness. Thermal
therapy, such as radiofrequency ablation (RFA), has been regarded
as an effective method to control tumor growth with an acceptable
morbidity rate. The main advantages of thermal therapy include 1)
minimally invasive with a high safety profile, 2) capability to
enable the focal tumor control, 3) favorable long-term survival
rates, and 4) it can be combined with other treatment approaches.
Despite the promising benefits provided by RFA, it does present
some clinical limitations. Due to the limited ablation volume and
heat shrink effect within the liver tissue, the ablation needle has
to be accurately placed within the tumor to create the desired
coagulation zone to encompass the entire target. However, manually
placing the RFA needle into the tumor within the dynamically moving
liver has been a long-standing challenge, even with
intra-procedural image guidance. Needle targeting error is mainly
caused by the respiration-induced movement of the liver, which can
be as large as 5.5 cm in the superior-inferior direction. To
address this issue, breath holding is typically required during the
placement of the needle to mitigate the motion. However, it can be
difficult for the patient to hold their breath due to compromised
lung capacity and the significant pain associated with the
radiofrequency ablation procedure.
BRIEF SUMMARY OF THE INVENTION
[0005] In one embodiment, the present invention provides a robotic
platform that enables accurate needle deployment in the dynamic
environment of the liver. The robotic platform has the following
features: (1) a lightweight, compact profile that can be mounted on
the patient directly; (2) provides precise ablation needle position
and orientation; and (3) stepwise needle insertion that mimics the
human insertion procedure.
[0006] In other embodiments, the present invention, provides a
respiratory compensated robot for liver cancer treatment by
providing an effective and efficient platform for the deployment of
radiofrequency ablation needles. The patient mounted robot consists
of a 4-DoF dual-stage cartesian platform used to control the
position of a 1-DoF needle insertion module. The robot is
biologically inspired as its grasp-insert-release technique of
needle deployment mimics current practices used by clinicians,
however with much greater accuracy. In comparison to previously
developed robots for needle interventions, the present invention is
capable of implementing an active motion compensation protocol that
minimizes any errors caused by the respiratory-induced motion of
the liver. The position of the compact and lightweight robot on the
patient can easily be adjusted to their comfort levels as compared
to other robots in the past that are either large in size or don't
interface with the patient at all, leading to potential targeting
errors.
[0007] In another embodiment, the present invention concerns a 5
degrees of freedom (DoF), patient mounted robot to perform
percutaneous needle interventions for medical procedures such as
the treatment of hepatocellular carcinoma.
[0008] In another embodiment, the present invention concerns a
robot that includes a dual cartesian platform with a custom 3D
printed active needle insertion module.
[0009] In another embodiment, the present invention concerns a
robot designed to be directly mounted onto the patient's
abdomen.
[0010] In another embodiment, the present invention concerns a
robot deploying a step-wise needle insertion module with an active
motion compensation protocol design to reduce targeting errors that
may be caused by the respiratory induced motion of the patient.
[0011] In another embodiment, the present invention concerns a
robot where, in free-space testing, the accuracy of the dual
cartesian platform was shown to have a mean error of 0.18.+-.0.18
mm in the x-direction, and a mean error of 0.32.+-.0.23 mm in the
y-direction.
[0012] In another embodiment, the present invention concerns a
robot where, in free-space testing, the needle insertion module has
an insertion accuracy of 0.64.+-.0.38 mm along the axis of the
needle, with good repeatability indicated by an average CV of
0.65%.
[0013] In another embodiment, the present invention concerns a
robot where the mean positional error is around 1.14.+-.0.30 mm,
and the mean orientational error is around
0.99.+-.0.36.degree..
[0014] In another embodiment, the present invention concerns a
robot where the mean positional error is around 1.22.+-.0.31 mm,
and the mean orientational error is around
1.16.+-.0.44.degree..
[0015] In another embodiment, the present invention concerns a
robot where the mean orientational error is around
1.66.+-.0.50.degree..
[0016] In another embodiment, the present invention concerns a
robot having a 66.3% to 69.6% improvement in positional accuracy
when an active motion compensation protocol is implemented.
[0017] In another embodiment, the present invention concerns a
robotic platform system that enables accurate needle deployment in
the dynamic environment of the liver comprising: (1) a lower stage
with a motorized cartesian carriage, (2) an upper stage, and (3) a
needle insertion module that connects both stages together.
[0018] In another embodiment, the present invention concerns a
robotic platform wherein the upper and lower stages have carriages
that can move in both the x and y directions.
[0019] In another embodiment, the present invention concerns a
robotic platform wherein the carriages both have spherical bearings
set into them that support the needle insertion module and by
changing the relative location between the bearings, the
orientation of the needle insertion module can be controlled about
the x and y axes.
[0020] In another embodiment, the present invention concerns a
robotic platform wherein the needle insertion module includes a
flexible fluidic actuator adapted to insert and retract a
needle.
[0021] In another embodiment, the present invention concerns a
robotic platform wherein the flexible fluidic actuator includes one
or more inflatable bellows adapted to move a needle linearly when
the bellows are inflated.
[0022] In another embodiment, the present invention concerns a
robotic platform wherein the flexible fluidic actuator includes one
or more diaphragms that hold a needle stationary within the robot
when inflated.
[0023] In another embodiment, the present invention concerns a
robotic platform wherein, when a needle is to be inserted further
into a patient, the one or more diaphragms are inflated prior to
the inflation of the bellows.
[0024] In another embodiment, the present invention concerns a
robotic platform wherein the upper and lower stages are
identical.
[0025] In another embodiment, the present invention concerns a
robotic platform further including a linear optical encoder and a
linear transmissive strip, the optical encoder and linear
transmissive strip adapted to provide the relative displacement of
the FFA.
[0026] In another embodiment, the present invention concerns a
robotic platform further including a gap between the needle and the
diaphragm and bellows when the diaphragms and bellows are
deflated.
[0027] In another embodiment, the present invention concerns a
robotic platform wherein the gap eliminates contact between the
needle and the bellows and the diaphragm when the bellows and the
diaphragm are deflated.
[0028] In another embodiment, the present invention concerns a
robotic platform wherein the needle does not contact the bellows
and the diaphragm when the bellows and the diaphragm are
deflated.
[0029] In another embodiment, the present invention concerns a
method for accurate needle deployment in the dynamic environment of
the liver comprising the steps of: providing a robotic platform
having (1) a lower stage with a motorized cartesian carriage, (2)
an upper stage, and (3) a needle insertion module that connects
both stages together; the upper and lower stages have carriages
that can move in both the x and y directions; the insertion module
includes of a fluidic actuator adapted to insert and retract a
needle; the flexible fluidic actuator includes one or more
inflatable bellows adapted to move a needle linearly when the
bellows are inflated and one or more diaphragms that hold a needle
stationary within the robot when inflated; when a needle is to be
inserted during a static phase of a liver, the one or more
diaphragms are inflated to grip the needle, then the one or more
bellows are inflated to insert the needle; and prior to a liver
transitioning out of a static phase, the one or more inflatable
diaphragms are deflated and then the one or more bellows are
deflated.
[0030] In another embodiment, the present invention concerns a
robotic platform wherein the needle is allowed to move freely
without interference within the fluid actuator when the needle is
not being inserted.
[0031] Additional objects and advantages of the invention will be
set forth in part in the description which follows, and in part
will be obvious from the description, or may be learned by practice
of the invention. The objects and advantages of the invention will
be realized and attained by means of the elements and combinations
particularly pointed out in the appended claims.
[0032] It is to be understood that both the foregoing general
description and the following detailed description are exemplary
and explanatory only and are not restrictive of the invention, as
claimed.
BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS
[0033] In the drawings, which are not necessarily drawn to scale,
like numerals may describe substantially similar components
throughout the several views. Like numerals having different letter
suffixes may represent different instances of substantially similar
components. The drawings illustrate generally, by way of example,
but not by way of limitation, a detailed description of certain
embodiments discussed in the present document.
[0034] FIG. 1 illustrates an embodiment of the present invention
concerning a liver ablation robot mounted on the patient within CT
scanner.
[0035] FIG. 2 illustrates a cartesian stage for use with an
embodiment of the present invention.
[0036] FIG. 3 illustrates the stages of the present invention
connected by a Z-frame.
[0037] FIG. 4 illustrates a flexible fluid actuator used with an
embodiment of the present invention.
[0038] FIG. 5 illustrates the flexible fluid actuator mounted to a
housing.
[0039] FIG. 6 is a cross-sectional view of a flexible fluid
actuator along with parameters used in a force estimation
analysis.
[0040] FIG. 7(a) illustrates the sequential step-like
grasp-insert-release method of actuation of the needle insertion
module as well as how the linear bellow will be depressurized to
the home position after the release step.
[0041] FIG. 7(b) shows the static and dynamic phases of an in-vivo
porcine liver.
[0042] FIG. 8 provides the coordinate frame assignment of the
patient mounted robot. Forward kinematics indicates that the robot
workspace (blue point cloud) is able to cover 70% of the liver.
[0043] FIG. 9 illustrates a clinical workflow of the robot for an
embodiment of the present invention.
[0044] FIG. 10 illustrates a desired path versus the actual tracked
path of the Cartesian stages wherein the dashed lines represent the
desired path while the dots show the data collected from the EM
tracker
[0045] FIGS. 11A, 11B and 11C illustrate the results of the FFA
characterization: (a) insertion force characterization results, (b)
retraction force characterization results, (c) experimental
friction force between the gripper and the needle as the pressure
input to the linear bellows is increased compared to the
theoretical peak friction force with the pressure input of 345 kPa
to the gripper mechanism.
[0046] FIG. 12 is a top view of the static targeting experiments.
The circles represent the desired target location, the stars show
the measured needle position in the static phantom, and the solid
stars show the measured needle position in the static ex-vivo
porcine liver.
[0047] FIG. 13 illustrates (a) Static phantom needle targeting
error vs. insertion depth (b) Static ex-vivo porcine liver needle
targeting error vs. insertion depth. In both tests, the needle
insertion angle was increased from 0.degree. to 15.degree. in
increments of 5.degree..
[0048] FIG. 14 is a comparison of dynamic phantom targeting
experiments with and without active motion compensation.
[0049] FIG. 15 is a comparison of ex-vivo porcine liver dynamic
targeting experiments with and without active motion
compensation
[0050] FIGS. 16A and 16B illustrate (a) Dynamic phantom needle
targeting error vs. insertion depth (b) Dynamic ex-vivo porcine
liver needle targeting error vs. insertion depth. In both
experiments, the needle insertion angle was increased from
0.degree. to 15.degree. in increments of 5.degree..
DETAILED DESCRIPTION OF THE INVENTION
[0051] Detailed embodiments of the present invention are disclosed
herein; however, it is to be understood that the disclosed
embodiments are merely exemplary of the invention, which may be
embodied in various forms. Therefore, specific structural and
functional details disclosed herein are not to be interpreted as
limiting, but merely as a representative basis for teaching one
skilled in the art to variously employ the present invention in
virtually any appropriately detailed method, structure or system.
Further, the terms and phrases used herein are not intended to be
limiting, but rather to provide an understandable description of
the invention.
[0052] In one embodiment, as shown in FIG. 1, the present invention
provides a robot 100 that is sized to fit within a CT scanner. The
robot consists of three major subsystems: (1) a lower stage 120
with a motorized cartesian carriage, (2) an identical upper stage
130, and (3) a needle insertion module 140 that connects stages 120
and 130 together. The upper and lower stages have, as shown in FIG.
2, carriages 200 that can move in both the x and y directions (2
DoF). The carriages both have spherical bearings 210 set into them
that support the needle insertion module. By changing the relative
location between bearings 210, the orientation of needle 300 (as
shown in FIG. 3) insertion module 140 can be controlled about the x
and y axes (2 DoF). The needle insertion module can both insert and
retract needle 300 through the use of a flexible fluidic actuator
310 (1 DoF) as shown in FIG. 3. Robot 100 is housed within a
3D-printed Z-frame 350 for purposes of locating the robot within
the CT-scanner. As further shown in FIG. 1, Z-frame 350 includes
adjustable straps 160 to attach robot 100 on the patient's abdomen
according to their comfort. The overall dimensions of the robot are
216 mm.times.210 mm.times.130 mm, ensuring a compact design such
that the patient can comfortably fit within the CT bore with the
robot.
[0053] As further shown in FIG. 2, upper and lower motorized
cartesian stages 120 and 130 include two fixed stepper motors 230
and 231 and provides a means of moving both axes independently or
simultaneously. The stationary motors have the added benefit of
reducing the number of moving parts, increasing the acceleration
capabilities of the robot, maintaining constant hardware gravity,
and allowing for a more compact design to be implemented. The
carriages are driven by a series of timing belts 240 attached to
the output shaft of four bipolar NEMA 11 stepper motors and the
acrylic base frame using friction-reducing pulleys. Limit switches
are used to set the home position of the carriages. Bearing
housings 250, shaft couplers and the carriages are FDM 3D printed
using the material acrylonitrile butadiene styrene (ABS). The
low-level control of the cartesian stages is done on a
microcontroller (ATmega 2560) with its supporting circuit. The two
stages are fixed parallel to each other and held in place by
support tabs attached to the fiducial registration Z-frame.
[0054] As shown in FIGS. 4-6, needle insertion module 140 consists
of two components that allow for 1-DoF needle insertion: (1) linear
bellow actuator 400 and (2) needle gripper 410. The mechanism is
also referred to as a flexible fluidic actuator (FFA) since both
components are actuated using pressurized air. The linear bellow
actuator consists of a toroidal geometry such that the hollow
center, which when pressurized with air, expands enabling the
translation or movement of needle 300 through the central axis of
the device. Under operation, needle 300 can be grasped using the
gripper mechanism 410. As shown in FIG. 6, gripper 410 consists of
one of more diaphragms 420A and 420B, that when inflated, expand to
grasp needle 300. By grasping the needle, then inflating the
bellows, needle insertion is achievable. FFA 450 may be 3D printed
in the material nylon-12 using selective laser sintering (SLS).
Additionally, as shown in FIG. 5, one or more safety tabs 500A and
50B may be used to restrict the linear translation of the mechanism
to <2 mm. The safety tabs are important to ensure safe operation
of the device to within only one full step of FFA 450 in the event
of a system failure. FFA 450 is housed within an FDM 3D printed
housing that consists of mechanical stop brackets 510A and 510B to
both restrict the linear displacement of the FFA as well as
constrain the displacement to pure translation. The safety tabs on
the FFA are designed to fit into these brackets and the step-size
of the mechanism can then be adjusted accordingly. Attached to the
housing is a linear optical encoder 550, while a linear
transmissive strip 470 (as shown in FIG. 4) is mounted to the
safety tab on the FFA, so that the relative displacement between
the housing and FFA could be measured to determine the insertion or
retraction depth of the needle. The FFA was controlled using two
proportional directional control valves with the control signal
supplied via a microcontroller and amplifier circuit.
[0055] Attached to the housing is an optical encoder (US-Digital
part no. EM1-1-500-N), and a linear transmissive strip (500 lines
per inch) is mounted to the safety tab on the FFA. This is done so
that the relative displacement between the housing and FFA could be
measured to determine the insertion or retraction depth of the
needle. The FFA was controlled using two proportional directional
control valves (Festo MPYE-5-M5-010-B) with the control signal
supplied via a microcontroller (ATmega 2560) and amplifier
circuit.
[0056] The robot is controlled using a Matlab implemented GUI that
communicates to the main microcontroller via serial bus. An active
motion compensation protocol is proposed for operating the patient
mounted robot. As shown in FIG. 7, in this protocol, the ablation
needle will be deployed towards the liver tumor during the
stationary phase of the respiration cycle which is when the liver
has the least motion and releases the needle when respiration
induced motion is significant. Specifically, prior to the static
phase of the liver, the needle is in a home position. When the
liver enters a static phase, the needle is first grasped by the
actuation of diaphragms 420A and 420B which occurs when the
diaphragms are pressurized. The needle is then inserted a
predetermined distance by the actuation of linear bellows 400.
Prior to the liver entering a dynamic phase the pressure being
supplied to the linear bellows and gripper is released, leaving the
needle in place in a patient. As shown in FIG. 8, this process is
repeated as the needle travels from the point of entry 810 to the
target area 820.
[0057] Using inflatable components has several advantages. First,
different size needles may be accommodated by the mechanism.
Second, the components may be completely retracted away from a
needle to create sufficient spacing or a gap between the FFA and
the needle thereby eliminating potential unwanted movement of the
needle during any movement of the liver. Thus, when the liver is in
the static phase, the needle may be inserted. When the liver is in
a dynamic phase, reducing or eliminating the contact between the
needle and the device prevents any unwanted repositioning of the
needle in the liver due to movement of the liver. In a preferred
embodiment, the needle does not touch the FFA at all during periods
when the needle is not being inserted.
[0058] With the CT Real-Time Position data feedback, the needle
insertion module must complete a full-step insertion when the liver
has negligible movement in its static phase and release the needle
when the liver undergoes significant movement during its dynamic
phase. The periodically moving target can be approximated to an
"immobilized" point with this "move-pause" needle insertion
protocol.
[0059] Several experiments were conducted in order to characterize
the performance of the robot: a) free-space accuracy evaluation of
the 4-DoF dual cartesian stages, b) free-space accuracy and force
characterization of the active needle insert unit, c) static
phantom and ex-vivo porcine targeting, d) dynamic phantom
targeting, and e) dynamic ex-vivo tissue targeting.
a) Free-space accuracy evaluation of the 4-DoF dual cartesian
stages revealed mean positional error in the x-direction of
0.18.+-.0.18 mm and a mean error in the y-direction of 0.32.+-.0.23
mm. b) Free-space accuracy evaluation of the active needle
insertion module revealed mean needle placement error of
0.64.+-.0.38 mm. Force characterization of the active needle
insertion module shows a peak frictional force between the gripper
mechanism and the needle of 22.6.+-.0.40 N indicating the force at
which the needle would slip. c) Static phantom targeting trials
showed a mean positional targeting accuracy of 1.14.+-.0.30 mm and
a mean orientational targeting accuracy of 0.99.+-.0.36.degree..
Static ex-vivo porcine liver targeting trials showed a mean
positional targeting accuracy of 1.22.+-.0.31 mm and a mean
orientational targeting accuracy of 1.16.+-.0.44.degree.. d) Using
the active motion compensation protocol, dynamic phantom targeting
trials showed a mean positional error of 1.69.+-.0.66 mm and a mean
orientational error of 1.66.+-.0.50.degree.. Without the active
motion compensation protocol implemented, dynamic phantom targeting
trials showed a mean positional error of 5.02.+-.2.35 mm and a mean
orientational error of 4.54.+-.1.40.degree.. e) Ex-vivo porcine
liver targeting trials were used to represent insertion into
biological tissue. Using the active motion compensation protocol,
dynamic porcine liver targeting trials showed a mean positional
error of 1.54.+-.0.55 mm and a mean orientational error of
1.68.+-.0.47.degree.. Without the active motion compensation
protocol implemented, dynamic porcine liver targeting trials showed
a mean positional error of 5.07.+-.2.44 mm and a mean orientational
error of 4.06.+-.1.45.degree..
[0060] Forward and Inverse Kinematics
[0061] The coordinate frames, F.sub.upper and F.sub.lower, for the
upper and lower stages are defined identically as being at the
center of the spherical bearings in the carriages when both stages
are in the homed position as seen in FIG. 8. The positions of the
center of the upper and lower carriages are denoted by
O.sub.upper=(x.sub.u, y.sub.u, z.sub.u), and O.sub.lower=(x.sub.l,
y.sub.l, z.sub.l) respectively. The x and y positions of the
carriages can be written in terms of the equations of motion
governing the mechanism that relate motor rotation .DELTA.A and
.DELTA.B to translation, given by
.DELTA.x=1/2(.DELTA.A+.DELTA.B) (1)
.DELTA.y=1/2(.DELTA.A+.DELTA.B) (2)
[0062] The forward kinematics of the robot takes the joint space
positions [x.sub.u,y.sub.u], [x.sub.l,y.sub.l] along with the
desired needle insertion depth, l, and solves for the position of
the tip of the needle, in addition to the needle axis vector. The
positions of the center of the carriages would then be defined as
O.sub.upper=(x.sub.u,y.sub.u,O) and
O.sub.lower=(x.sub.l,y.sub.l,-H), where H is the constant distance
maintained between the two parallel stages of the cartesian
platform. The unit vector, {circumflex over (N)}, defining the axis
of the needle could then be given by the normalized vector of the
difference between the positions of the center of the upper and
lower carriages. Using this needle axis vector, for the position of
the of the tip of the needle, T.sub.desired using the insertion
depth may be solved. The equations defining {circumflex over (N)}
and T.sub.desired are
N ^ = O l .times. o .times. w .times. e .times. r - O u .times. p
.times. p .times. e .times. r O l .times. o .times. w .times. e
.times. r - O u .times. p .times. p .times. e .times. r ( 3 ) T d
.times. e .times. s .times. i .times. r .times. e .times. d = O l
.times. o .times. w .times. e .times. r - l .times. N ^ ( 4 )
##EQU00001##
[0063] The robot workspace was simulated in Matlab across the
achievable translations in the x- and y-directions of the cartesian
stages described in Table 1.
TABLE-US-00001 TABLE 1 FIG. 1. CAD model of the liver ablation
robot mounted on the patient within CT scanner. Robot Properties
Based on Design and Workspace Analysis Robot Dimensions 216 mm
.times. 210 mm .times. 130 mm Active DoF 5 x-y Displacement 75 mm
.times. 70 mm Insertion Depth Limited by Needle Length Orientation
about x and y axes .+-.15.degree. Weight 2.17 kg
[0064] The markings on FIG. 8 show the workspace of the robot and
it is overlaid with an average adult human liver. Comparing the
volume of the workspace to that of an average adult human liver,
the robot can reach 70% of the volume of an average adult human
liver when placed directly above it. Note that the 70% value is
just a comparison of the robot workspace to the total size of the
average liver
[0065] The inverse kinematics of the robot is used to solve for the
joint space positions x.sub.u, y.sub.u, x.sub.l and y.sub.l along
with the required needle insertion depth, l, given the desired
needle tip position, T.sub.desired=(T.sub.x, T.sub.y, T.sub.z), and
the needle entry point, P.sub.entry=(P.sub.x, P.sub.y, P.sub.z),
within the global coordinate frame. The needle vector axis can be
defined by
N=T.sub.desired-P.sub.entry (5)
[0066] Using N=(N.sub.x, N.sub.y, N.sub.z), we can then solve for
the positions of the center of the carriages needed to generate
this needle axis vector given by,
x u = T x - T z + H N z .times. N x ( 6 ) y u = T y - T z + H N z
.times. N y ( 7 ) x l = T x - T z N z .times. N x ( 8 ) y l = T y -
T z N z .times. N y ( 9 ) ##EQU00002##
[0067] The insertion depth of the needle, l, is solved for by
finding the Euclidean distance between the desired entry point and
the location of the desired target,
l=.parallel.T.sub.desired-P.sub.entry.parallel. (10)
[0068] In the clinical workflow the point P.sub.entry will be
defined by the clinician based on pre-operative imaging. The robot
forward and inverse kinematics were implemented in a custom Matlab
GUI, whereby T.sub.desired and P.sub.entry could be input and the
coordinates of the upper and lower carriages are solved for to
generate the appropriate needle vector axis. These joint space
positions in addition to the needle insertion depth are then
communicated to the robot over a serial bus to the low-level
microcontroller such that the robot can be aligned to the desired
location.
[0069] Active Motion Compensation Protocol
[0070] An active motion compensation protocol is proposed for
operating the patient mounted robot. In this protocol, the ablation
needle will be automatically deployed towards the liver tumor
during the stationary phase of the respiration cycle. The
respiratory cycle of the patient will be gated using the GE D690
PET/CT scanner which uses the Varian CT Real-Time Position
Management system (Varian Medical Systems, Palo Alto, Calif.). The
Real-Time Position Management system consists of reflectors
attached to an external marker placed on the patient's abdomen. The
marker motion reflects the breathing pattern of the patient and can
be captured by an external camera at a frequency of 30 Hz to obtain
a surrogate respiratory signal. The system is able to track the
real-time position data even when the respiratory rate changes
suddenly. The needle insertion module must complete a full-step
insertion when the liver has negligible movement in its static
phase and release the needle when the liver undergoes significant
movement during its dynamic phase. In order to achieve this
protocol, the insertion sequence is divided into four sequential
steps: grasp, insert, release, and home. FIG. 7(a) shows the finite
element method (FEM) simulation of the FFA motion subject to the
pressure input. Retracting the needle can be achieved by
alternating the sequence to power the FFA. Note that the operation
process must be completed within about 2 s to cover just the static
phase as seen in FIG. 7(b). The current prototype of the needle
insertion module is capable of inserting the needle at a maximum
speed of 1.5 mm/s.
[0071] Robotic RFA Clinical Workflow
[0072] In a clinical workflow, the patient will be positioned in a
PET/CT scanner and the robot will be fixed to the patient's abdomen
using adjustable straps. An initial CT scan will be performed to
register the robot to the CT scanner using a point-based
registration method. A second CT scan will be used to identify the
desired target location, which will be used by the radiologist to
determine an appropriate needle entry point. The robot will
automatically align the needle insertion module to the desired
target and the needle can then be progressed towards the ablation
target. During the needle insertion process, the Varian CT
Real-Time Position Management system will continuously track the
patient's respiratory cycle. According to the proposed motion
compensation protocol, the needle is only advanced when the liver
has the least amount of motion (static phase, see FIG. 7(b). Once
the needle is placed at the target location, a confirmation scan
will be conducted to verify the needle location, and ablation
therapy will then be performed. The detailed workflow is shown in
FIG. 9.
[0073] Force Modeling of the Needle Insertion Module
[0074] Force modeling of the FFA may be used to estimate the force
the FFA is capable of inserting a needle with. The axial force of
the linear bellows, F.sub.A, is given by,
F A = .pi. .function. ( R + h ) 2 .times. P ( 11 ) where x = ( 1 -
.rho. 2 ) .times. ( 1 - .rho. 4 + 4 .times. .rho. 2 .times. .times.
ln .times. .times. .rho. ) 4 .times. ( 1 - .rho. 2 + 2 .times.
.rho. .times. .times. ln .times. .times. .rho. ) .times. ( 1 -
.rho. 2 - 2 .times. .rho. .times. .times. ln .times. .times. .rho.
) ( 12 ) .rho. = R - h R + h ( 13 ) ##EQU00003##
[0075] The parameter R defines the mean radius of the bellow, and h
is half the wave height of a corrugation, as depicted in FIG. 6. In
addition to this axial force caused by the linear bellows, the
force of friction that the gripper is holding the needle with plays
a key role in determining the needle insertion force and should be
considered. From a review of the principles from mechanics of
materials, the deflection of a thin circular plate with clamped
outer edges and a free inner edge is given by
y max = P .times. a 4 6 .times. 4 .times. D ( 14 ) ##EQU00004##
[0076] where P is a uniformly distributed pressure over the surface
area of the diaphragm, a is the radius of the diaphragm, and D is
the flexural rigidity defined in (15) as
D = E .times. .times. t 3 1 .times. 2 .times. ( 1 - v 2 ) ( 15 )
##EQU00005##
[0077] where E and v are the material properties Young's modulus
and Poisson's ratio, respectively, and t is the thickness of the
diaphragm. Again, from mechanics of materials, the normal force,
F.sub.N, acting on the needle due to the deflection of the
diaphragms may be derived. The relationship between the normal
force and the displacement at the center of the diaphragm is given
by,
y max = F N .times. a 2 1 .times. 6 .times. .pi. .times. D ( 16 )
##EQU00006##
[0078] Substituting (14) and (15) into (16) provides the
relationship between the input internal pressure of the diaphragm
chamber and the normal force generated between the diaphragms and
the needle. A scaling factor of 2 is included since the normal
force is being applied to both sides of the needle. This
relationship is given by,
F N = .pi. 2 .times. P .times. a 2 ( 17 ) ##EQU00007##
[0079] Using this normal force to pressure relationship, it is
possible to estimate the force of friction holding the needle in
place by multiplying the normal force by the coefficient of
friction, 0.35, of the material. For an input pressure of 345 kPa,
an estimated 24.6 N of friction force will be used to clamp onto
the needle. This force is considered as more than sufficient for
percutaneous needle interventions.
[0080] Experiments and Results
[0081] Several experiments were conducted to characterize the
performance of the robot: a) free-space accuracy evaluation of the
4-DoF dual cartesian stages, b) free-space accuracy and
repeatability of the needle insertion module, c) force
characterization of the needle insertion module, d) static phantom
and static ex-vivo porcine liver tissue targeting, e) dynamic
phantom targeting, and f) dynamic ex-vivo porcine liver tissue
targeting.
[0082] Accuracy Analysis of the Dual Cartesian Stages
[0083] To characterize the accuracy of the dual cartesian stages, a
free-space analysis was conducted. This was done using the Aurora
electromagnetic (EM) tracking system (NDI Medical, Ontario, Canada)
with resolution of 0.5 mm. The two carriages were linked together,
allowing the central axis between the two carriages to be more
easily identified. A 5-DoF EM sensor was used to track the position
of the central axis between the carriages. Coordinate registration
was done to be able to track the sensor within the robot reference
frame using the point-based registration method. The reference
frame registration was taken after initially homing both stages of
the robot. From there, a desired path was sent to the robot via the
Matlab GUI and the real time position was tracked. The desired
paths consisted of four squares of length 10 mm, 20 mm, 30 mm, and
40 mm, all created by streamlining the coordinates of the corners
of the square to the GUI as seen in FIG. 10. The mean error across
three experiments for each desired path was measured at the corners
of the squares in both the x- and y-directions. The mean error in
the x-direction is 0.18.+-.0.18 mm and the mean error in the
y-direction is 0.32.+-.0.23 mm. These free-space experimental
results were taken as validation for sufficient accuracy of the
dual cartesian system to potentially be used for precise needle
placement operations.
[0084] Accuracy and Repeatability of Needle Insertion Module
[0085] To characterize the accuracy of the needle insertion module,
the EM tracker was again used to perform insertion in free-space.
An EM sensor was attached to the tip of a needle and a custom 3D
printed bracket was used for coordinate registration. The needle
used was an 18-gauge needle with diamond shaped tip. Eight
insertion depths of 15 mm, 30 mm, 45 mm, 60 mm, 75 mm, 90 mm, 105
mm and 120 mm were sent to the needle insertion module from the GUI
for a total of three iterations for each depth. The error was
considered to be the difference between the measured insertion
depth and the desired depth. The mean error across all 24
experiments was found to be 0.64.+-.0.38 mm (see Table 2 for the
error results at each insertion depth). Note that the error does
not accumulate as the needle insertion depth is increased due to
the step-wise operation of the needle insertion module. This
step-wise insertion ensures that the needle insertion error is
within one full step-size of the FFA (1.5 mm). The repeatability of
the needle insertion module was quantified by evaluating the
Coefficient of Variation (CV) [41, 54]. The CV is expressed as a
percentage, and the lower the percentage, the better the
repeatability is. The results of the needle insertion experiment
are presented in Table 2.
TABLE-US-00002 TABLE 2 Needle insertion accuracy and repeatability
Desired Mean STD of the Insertion Depth Measured |Mean Error|
Measured (mm) Depth (mm) (mm) Depth (mm) CV % 15 15.94 0.94 0.10
0.62 30 30.49 0.49 0.19 0.61 45 44.97 0.13 0.25 0.55 60 60.26 0.84
0.80 1.33 75 76.12 1.12 0.51 0.67 90 90.40 0.56 0.49 0.54 105
105.17 0.56 0.54 0.51 120 120.16 0.36 0.43 0.36 Mean 0.64 0.38
0.65
[0086] Force Characterization of Needle Insertion Module
[0087] In addition to the insertion testing accuracy in free-space,
the forces generated by the FFA were also characterized. The
insertion force due to pressurizing the linear bellows was found by
placing a force sensor (Vernier Go Direct.RTM. Force and
Acceleration Sensor) with a flat plate adapter directly against the
FFA. The pressure inside the linear bellows was slowly increased
from 0 kPa to 110 kPa, corresponding to the maximum force
measurable by the force sensor. As the pressure was increased, the
insertion force was recorded at intervals of 17.5 kPa (2.5 psi). A
custom 3D printed bracket was designed to attach the safety tabs on
the FFA to the force sensor. It was configured such that the linear
bellows could be pressurized and allow the pulling, retraction
force of the bellows to be measured as they were depressurized. As
the pressure was decreased, the corresponding retraction force was
recorded for the same intervals as previously mentioned. The
results of the linear bellows force characterization were compared
to the predicted forces calculated using (11)-(13) set forth above
as seen in FIGS. 11(a) and (b). A mean error of 2.57.+-.0.6 N was
found for the insertion force experiment and a mean error of
4.04.+-.1.39 N was found for the retraction force experiment. These
errors can be largely attributed to manufacturing imperfections in
the FFA.
[0088] The linear bellows are capable of producing significantly
more force than the gripping diaphragms at the same pressure input,
therefore it is expected that the needle will slip between the
diaphragms at a certain point. To characterize this frictional
force, a needle was placed between the grippers at 345 kPa and the
force sensor then positioned at the tip of the needle. The pressure
inside the linear bellows was then increased and the force at the
tip of the needle was monitored. A noticeable peak force was
achieved which corresponded to the needle beginning to slip due to
the change from static friction to dynamic friction, as seen in
FIG. 11(c). Based on the calculations above, a peak force of 24.6 N
was expected, while the experimental results indicate a mean peak
force of 22.6.+-.0.40 N across three trials. This difference of
8.4% may be as a result of manufacturing imperfections and
environmental conditions affecting the coefficient of friction
between the two materials. The linear bellows force
characterization and the friction force analysis both indicate more
than sufficient force is achievable to perform percutaneous liver
interventions, where a peak force of about 6 N was recorded for the
percutaneous interventions based on previously mentioned
research.
[0089] Robot Targeting Test: Static Phantom and Ex-Vivo Porcine
Liver Trial
[0090] A 5% by volume agar gelatin phantom was created to mimic
soft liver tissue. The experiment was conducted by selecting 24
arbitrary points, grouped into four groups of 6 points, within the
robot's workspace and using their xy-location as the desired
target. For the first group of six points, the desired insertion
depth was increased from 15 to 90 mm in 15 mm increments across the
points and the needle orientation was set to 0.degree.. This was
then repeated for the remaining three groups however the needle
orientation about the x-axis was increased in 5.degree. increments
across the groups. Using these values as the desired input to the
robot GUI, the needle was placed and automatically deployed to the
phantom via the needle insertion module. The final location of the
needle was recorded by the EM tracker. Each targeting experiment
was conducted three times and the mean values were used in
quantifying the accuracy of the robot. The positional error was
defined to be the Euclidean distance between the desired target and
the measured location. Additionally, the orientational error is
defined by the difference between the desired input angle and the
measured angle. The results for the static phantom experiment show
that there is a mean positional error of 1.14.+-.0.30 mm and an
orientational error of 0.99.degree..+-.0.36.degree..
[0091] The same points and grouping scheme was used to repeat the
experiment in a static ex-vivo porcine liver except the insertion
depth increased from 10 to 60 mm in increments of 10 mm, where 60
mm was the maximum thickness of the porcine liver sample used.
Porcine liver has often been used to mimic an environment for
testing needle insertion devices in place of human tissue. The
porcine liver was acquired fresh from a local meat supplier, and
prior to testing the liver was allowed to come to room temperature.
The results show that there is a mean positional error of
1.22.+-.0.31 mm and an orientational error of
1.16.degree..+-.0.44.degree.. A top view of the static targeting
experiments can be seen in FIG. 12. There is a slight increase in
positional targeting error of the needle with respect to the
insertion depth in both the static phantom and ex-vivo liver
trials. This can be seen in FIG. 13 where there is a general
increasing trend in the positional error as the insertion depth is
increased. The targeting error, however, does not show any
significant statistical relation to the increasing inclination of
the needle.
[0092] Robot Targeting: Dynamic Phantom Trial
[0093] To validate the robot's targeting performance under the
dynamic conditions of the liver, a dynamic motion platform was
developed to create a relative displacement between the robot and
the phantom. The motion platform consists of two stepper motor
driven linear rails fixed to one another perpendicularly in the
horizontal plane. A two-dimensional dynamic phantom was chosen
rather than considering three-dimensional motion for two primary
reasons. It has been found that liver motion in the
inferior-superior (I-S) and left-right (L-R) is dominant in
comparison to the motion generated in the posterior-anterior (P-A)
direction (.about.1.2 mm).
[0094] A two-dimensional dynamic motion platform also greatly
simplified the robotic motion platform, especially during the
prototype characterization period. Between every two steps of the
needle insertion process, the motion platform moves the phantom 10
mm in the x-direction and 5 mm in the y-direction to simulate the
respiratory induced motion of the liver in the I-S direction and
L-R direction, respectively. While the phantom is in motion, the
needle insertion module releases the needle to allow it to move
with the phantom freely. Once the platform has returned to its
original position, the needle insertion module takes another two
full insertion steps and the process is repeated until the needle
has reached its final target.
[0095] The stepwise "move-pause" insertion protocol is inspired by
the manual insertion procedure, where the clinician typically
inserts the needle when the liver has the least motion and releases
the needle when respiration motion is significant. Twenty-four
targets were selected by arbitrarily inserting the 6-DoF EM
tracking probe (NDI Medical, Ontario, Canada) into the phantom. The
location of the target, the tip of the EM tracking probe, was then
converted into the robot frame and used as an input for the GUI.
Similarly to the static phantom trial, the 24 points were grouped
into 4 groups of 6 and the desired insertion depth of the needle
was user controlled from 15 to 90 mm in 15 mm increments while the
needle orientation across each group was increased from 0.degree.
to 15.degree. in increments of 5.degree.. Three insertion trials
were conducted for each target and the error for these dynamic
targeting experiments is defined to be the same as in the static
experiments. The results of the dynamic targeting experiments in
phantom indicate a mean positional error of 1.69.+-.0.66 mm and a
mean orientational error of 1.66.+-.0.50.degree..
[0096] The dynamic targeting experiments were repeated, however,
this time no motion compensation was considered. The synchronicity
between the needle insertion module and the motion platform was
disabled so that there was no consideration of the location of the
moving target with respect to the insertion of the needle. The
needle was inserted with constant speed until it reached its final
target, the motion platform was simultaneously stopped, and the
final needle position was recorded. The results of this dynamic
targeting experiment with no motion compensation indicates a mean
positional error of 5.02.+-.2.35 mm and a mean orientational error
of 4.54.+-.1.40.degree.. A comparison between the dynamic phantom
targeting experiments with active motion compensation and without
motion compensation is shown in FIG. 14.
[0097] Robot Targeting: Dynamic Ex-Vivo Porcine Liver Trial
[0098] To simulate targeting in biological tissue, an ex-vivo
porcine liver sample was used in place of the agar phantom. The
same workflow used in the dynamic phantom targeting experiments was
employed here for a total of 24 targets. With the motion
compensation protocol implemented, the results indicate a mean
positional error of 1.54.+-.0.55 mm and a mean orientational error
of 1.68.+-.0.47.degree.. The experiments were repeated without the
motion compensation protocol implemented and the results show a
mean positional error of 5.07.+-.2.44 mm and a mean orientational
error of 4.06.+-.1.45.degree.. A comparison of these results is
presented in FIG. 15. Similar to the static experiments, a slight
increase in positional targeting error was observed in both the
phantom and ex-vivo liver trials, more noticeably in the ex-vivo
liver as seen in FIG. 16. Again, no noticeable statistical
correlation could be made between targeting error and increasing
inclination of the needle.
[0099] While the foregoing written description enables one of
ordinary skill to make and use what is considered presently to be
the best mode thereof, those of ordinary skill will understand and
appreciate the existence of variations, combinations, and
equivalents of the specific embodiment, method, and examples
herein. The disclosure should therefore not be limited by the
above-described embodiments, methods, and examples, but by all
embodiments and methods within the scope and spirit of the
disclosure.
* * * * *